Statin-Associated Memory Loss

Discussion

The cardiovascular benefits of statins are well established, and they reduce the risk of death by 14-28% in specific populations.^[17] However, the effects of these agents on the human brain are not as well established. The more lipid-soluble the statin, the greater propensity it has to cross the blood-brain barrier and affect the central nervous system.^[18] According to some reports,^[1,8] simvastatin is the most lipophilic drug in its class; pravastatin is the least lipophilic. Some reports state that atorvastatin does not cross the blood-brain barrier, whereas others state that the drug has intermediate lipophilicity.^[1,8] If the effects of statins on memory and dementia are mediated directly at the neural level, then one would predict that lipophilicity would be correlated with both the protective and the adverse neural effects associated with statins.

As reported above, there is great interest in developing statins as a treatment for dementia and as an agent for prevention of dementia in healthy elderly individuals. Our study was initiated because of this interest, along with reports in the lay press and consumer forums regarding statin-associated memory loss. Because an initial literature search yielded few such published cases, we searched MedWatch to study the clinical characteristics of reported statin-associated memory loss. As such, this is the first scientific endeavor to systematically analyze a relatively large series of cases (60) of documented memory loss associated with statins. In our series, more than half the reports were from consumers. The main symptom appeared to be short-term memory loss that occurred a few months after the start of statin therapy or after a dosage increase.

Global or partial amnesia was also reported. All but one of the reports identified using our search criteria involved atorvastatin or simvastatin. In some cases, memory loss resolved completely after discontinuation of the statin; in others it did not. In four cases, rechallenge appeared to reproduce memory loss. Most reports contained only subjective information. No dose-effect relationship could be demonstrated because of insufficient data. Similarly, no relationship between any laboratory parameters (e.g., lipid levels) and memory could be demonstrated because of the lack of such data in the reports. Confounding concomitant variables (e.g., transient ischemic attacks or other drugs that may cause memory loss) were present in some but not all reports. Information on medical history, or concomitant diseases or drugs, was absent or incomplete in most reports. Because many of the reports were from consumers, there is some reason to be cautious about the accuracy of reported medical information. However, lay persons can accurately notice short-term cognitive impairments caused by drugs such as alcohol or benzodiazepines.

The information in the 60 reports of spontaneous adverse events we analyzed do not permit conclusive judgments about causality. The high background rate of memory loss in this population due to aging and vascular risk factors could lead to detection bias. Likewise, the absence of objective memory tests or lipid data made it difficult to confirm the memory change and establish a dose-severity relationship or links with cholesterol levels. Cognitively high-functioning individuals may often experience subtle memory loss that they can accurately report but is often not detectable even with neuropsychological testing.

A review of the randomized clinical trial data does not shed additional light on this issue. Many large published cardiovascular trials of statins did not use a formal neuropsychological test battery to assess for memory outcomes, but used quality-of-life measures, which are not as sensitive or specific for memory. There may be a number of other statin studies that collected such memory data but are not published.

In one published placebo-controlled trial of lovastatin, the placebo appeared to have greater consistent cognitive benefits than lovastatin on some measures over a 6-month period.^[11] It has been suggested that a very low cholesterol level in neuronal membranes may decrease cognitive function, whereas others have disagreed with this hypothesis.^[19] More important, one large prospective study (PROSPER) found no significant benefit for pravastatin regarding stroke risk, cognition, or disability.^[10] As cognitive measures, this study used a global measure of cognition as well as measures of memory and speed of per-formance. Although cognition was a secondary outcome, it can be reasonably concluded from this study that pravastatin over a 3-year period does not provide any cognitive benefits.^[10]

A review of observational studies in the literature^[2,3,4,5,6] yielded several large case-control studies that suggest a substantial protective effect of statins on lowering the risk of dementia. These studies used statistical adjustments or case-control matching to account for potential biases (e.g., concomitant illnesses or other baseline differences). The fact that these findings were present in diverse population samples across five studies in the United States and the United Kingdom is a strength. Because any protective effect against dementia must be associated with a slower rate of memory loss, these data argue against an adverse effect of statins on memory.

The observational studies^[2,3,4,5,6] may have been limited by their retrospective design, inaccurate diagnoses of dementia, and indication biases. These studies did not distinguish type of statin, dosage, or duration of therapy. In the absence of dosage or duration effects and retrospective data analyses, their findings are at best preliminary. Patients with dementia, lower economic status, or less education may be less likely to be prescribed statins than those without dementia; hence, there could be a cohort bias as well.^[20]

Our report does not estimate frequency or prevalence of statin-associated memory loss since MedWatch case reports are likely to underestimate the true rates of adverse events. Approximately 2% of all reports or statin-associated adverse events in MedWatch appear to have a cognitive or amnestic identifier. Likewise, the prescribing information for pravastatin lists memory loss as an adverse event in less than 1% of patients,^[21] whereas that for simvastatin reports memory loss as a class effect.^[22] The prescribing information for atorvastatin lists amnesia as an adverse event in less than 2% of patients^[23] ( Table 3 ). Hence, these are likely to be rare.

According to a written communication from the manufacturer of atorvastatin (O. J. Lopena, Pharm.D., Pfizer Inc., written communication, 2002), amnesia during clinical trials occurred in 7 (0.3%) of 2502 patients receiving atorvastatin and in 2 (0.3%) of 742 subjects during comparative trials with other statins (lovastatin, pravastatin, simvastatin). Abnormal thinking was reported in 4 (0.2%) of 2502 patients receiving atorvastatin therapy and in none of 742 patients receiving other statins (lovastatin, pravastatin, simvastatin).^[24] None of these data may provide a true estimate of the frequency or prevalence since memory loss was not specifically studied in many of these trials. Incidence studies of other adverse events (e.g., sexual dysfunction with antidepressants) have shown that spontaneous reporting may underestimate prevalence, in some cases by 5-20-fold.

Our review of MedWatch reports and the literature provided no definite evidence that a particular statin is more likely than others to be associated with cognitive adverse effects or benefits. Because of limited resources, we examined only three statins. There appear to be fewer cases of memory loss associated with pravastatin in the 60 case reports we selected, although this may simply reflect our selection criteria or less frequent use of the drug.

There are many possible mechanisms by which statins may benefit or impair cognition. Statins may increase endothelial nitric oxide synthase and reduce endothelin-1, thereby increasing cerebral blood flow.^[25] The antioxidant, antiinflammatory, and platelet effects of statins may also play a role in neuroprotection ( Table 4 ).^[25,26,27]

In vitro and in vivo experimental animal studies indicate that statins reduce amyloidogenesis,^[1,26] although a human study found no effects of simvastatin on amyloid.^[14] Cholesterol is critically involved in cell membrane integrity and function; thus, one could speculate that cognitive function may worsen if excess inhibition of cholesterol synthesis reduces neuronal membrane levels. Optimal harnessing of the cardiac and neural benefits of statins remains a desirable goal.

Cite this: Statin-Associated Memory Loss: Analysis of 60 Case Reports and Review of the Literature - Medscape - Jul 01, 2003.

Table 1. Published Prospective, Randomized, Controlled Studies of Effects of Statins or Diet on Cognition
Treatment	No. of Patients	Age Range (yrs)	Results
Simvastatin^[9]	20,536	40-80	No statistically significant difference in cognition between statin and placebo groups.
Pravastatin^[10]	2891 (pravastatin) 2913 (placebo)	70-82	No statistically significant effects on MMSE, word recall, or performance time.
Lovastatin^[11]	98 (lovastatin) 96 (placebo)	24-60	Placebo group improved from baseline on all neuropsychological tests (p<0.04); lovastatin group improved from baseline only on memory recall test (p=0.03).
Simvastatin, pravastatin^[12] (crossover study)	36	40-60	No statistically significant effects on word recall test, profile of mood states, choice reaction time test, symbol substitution, or drowsiness rating.
Simvastatin, pravastatin^[13] (crossover study)	25	20-31	No statistically significant effects on EEG potential, mood, sleep, or cognitive performance.
Simvastatin^[14]	24 (simvastatin) 20 (placebo)	59-77	No statistically significant effects on CSF levels, AB40, or AB42.
Low-fat diet, low-cholesterol diet^[15]	52 (low-fat diet) 53 (low-cholesterol diet) 50 (control)	41-65	At 12 wks, results of the sustained attention task differed significantly (p<0.001) in groups with reduced cholesterol.

Table 2. Observational Studies of Effects of Statins on Cognition
Design	Statins Prescribed	Treatment Group, No. of Patients	Age (yrs)	Results
Epidemiologic, retrospective nested case-control study^[2]	Atorvastatin, cerivastatin, fluvastatin, pravastatin, simvastatin	Group 1: Dementia, 284, Statins, 13 Group 2: Controls, 1080, Statins, 104	≥ 50	Relative risk of dementia was reduced 0.29 in patients taking statins (95% CI 0.13-0.63, p=0.002).
Retrospective, multicenter, epidemiologic, cross-sectional analysis^[3]	Lovastatin, pravastatin, simvastatin	Lovastatin, 4180 Pravastatin, 2326 Simvastatin, 3580	≥ 60	Prevalence of AD was 60-73% lower in patients receiving lovastatin or pravastatin than in the entire cohort (p<0.001).
Retrospective cohort case-control study^[4]	Not specified	Statins, 57 Controls, 2248	≥ 65	Statins were associated with a lower risk of dementia in patients aged < 80 years (OR 0.26, 95% CI 0.08-0.88).
Multicenter secondary analysis^[5]	Simvastatin, atorvastatin, pravastatin, lovastatin, fluvastatin	Statins, 583 Controls, 454	< 80	Modified MMSE scores were higher in the statin group (93.7) vs the control group (92.7) (p=0.02).
Retrospective study of postmenopausal women^[6]	Not specified	Statins, 113 Controls, 542	52-98	Prevalence of dementia-AD was decreased (p<0.05) and MMSE scores were higher in the statin vs the control group (p=0.025).

Table 3. Comparison of Prescribing Information
Statin	Recommended Dosage (mg/day)	Frequency of Cognitive-Related Adverse Events
Atorvastatin^[23]	10-80	< 2% (amnesia)
Pravastatin^[21]	10-80	< 1% (memory impairment)
Simvastatin^[22]	5-80	Not specified

Table 4. Potential Mechanisms by Which Statins May Affect Brain Function
Effect	Mechanism
Platelet activity	platelet aggregation and deposition onto damaged vessel walls
Thrombin generation	thrombus generation generation of thrombin cleavage peptides
Nitric oxide formation	cerebral blood flow toxic production of nitric oxide
Antiinflammatory effects	-amyloid peptides A 42 and A 40 formation of proinflammatory isoprenoids expression of adhesion molecules elaboration of potentially damaging cytokines from macrophages during cerebral ischemia
Antioxidant effects	free radical injury and lipoprotein oxidation
Inhibition of cholesterol synthesis	Can be good or bad. May interfere with neuronal function if statin is lipophilic; cholesterol essential for membrane integrity.